Asymmetric mesh NoC topologies

1. A method of interconnecting blocks on a floorplan for a Network on Chip (NoC) topology, the method comprising:

receiving at least physical dimensions of a plurality of on chip blocks on the floorplan

receiving physical dimensions of the floorplan and relative placement information for placing the plurality of on chip blocks on the floorplan;

receiving porosity information for the plurality of on chip blocks on the floorplan; and

synthesizing, using a computer, an interconnection network and a plan for a physical layout of the interconnection network for the plurality of on chip blocks on the floorplan based on the received physical dimensions and porosity information and required chip functionality,

wherein the plurality of on chip blocks comprise on chip blocks of heterogeneous dimensions.

2. The method of claim 1 , wherein ones of a plurality of routers are retained on a mesh reference grid only at locations where message packets communicated between the plurality of on chip blocks change direction and at locations where message packets exit to host blocks of the plurality of on chip blocks.

3. The method of claim 2 , further comprising:

determining dimensions of the mesh reference grid based on the physical dimensions of the floorplan, the physical dimensions of the plurality of on chip blocks, the relative placement information of the plurality of on chip blocks, and connection port requirements of a block of the plurality of on chip blocks;

placing the plurality of on chip blocks on the mesh reference grid based on the relative placement information;

determining ones of a plurality of routers on the reference grid that are unusable based on the porosity information for the plurality of on chip blocks;

removing the unusable ones of the plurality of routers;

removing redundant series-connected ones of the plurality of routers and merging physical links associated with the removed redundant series-connected ones of the plurality of routers into a single physical link; and

replacing the single physical link with one or more pipeline register stages or one or more buffer stages.

4. The method of claim 3 further comprising:

mapping inter-block communication over the remaining ones of the plurality of routers and associated physical links to the remaining ones of the plurality of routers.

5. The method of claim 2 , wherein at least one of the plurality of on chip blocks comprises multiple ports connected to one of (a) multiple routers of the plurality of routers and (b) multiple ports of the one router of the plurality of routers.

6. The method of claim 2 , wherein at least one of the plurality of routers comprises multiple ports, each of the multiple ports connected to a different one of the plurality of on chip blocks.

7. A non-transitory computer readable medium storing instructions for implementing a method, the instructions comprising:

receiving at least physical dimensions of a plurality of on chip blocks on a floorplan for a Network on Chip (NoC) topology;

receiving physical dimensions of the floorplan and relative placement information for placing the plurality of on chip blocks on the floorplan;

receiving porosity information for the plurality of on chip blocks; and

synthesizing an interconnection network and a plan for a physical layout of the interconnection network for the plurality of on chip blocks on the floorplan based on the received physical dimensions and porosity information and required chip functionality, wherein the plurality of on chip blocks comprise blocks of heterogeneous dimensions.

8. The non-transitory computer readable medium storing instructions for implementing the method of claim 7 , wherein ones of a plurality of routers are retained on a mesh reference grid only at locations where message packets communicated between the plurality of on chip blocks change direction and at locations where message packets exit to host blocks of the plurality of on chip blocks.

9. The non-transitory computer readable medium storing instructions for implementing a method of claim 8 , wherein the instructions further comprise:

determining dimensions of the mesh reference grid based on the physical dimensions of the floorplan, the physical dimensions of the plurality of on chip blocks, the relative placement information of the plurality of on chip blocks, and connection port requirements of a host block of the plurality of on chip blocks;

placing the plurality of on chip blocks on the mesh reference grid based on the relative placement information;

determining ones of a plurality of routers that are unusable based on the porosity information for the plurality of on chip blocks;

removing the unusable ones of the plurality of routers;

removing redundant series-connected ones of the plurality of routers and merging the physical links associated with the removed redundant series-connected ones of the plurality of routers into a single physical link; and

replacing the single physical link with one or more pipeline register stages or one or more buffer stages.

10. The non-transitory computer readable medium storing instructions for implementing a method of claim 9 , wherein the instructions further comprise:

mapping inter-block communication over the remaining ones of the plurality of routers and associated physical links to the remaining ones of the plurality of routers.

11. The non-transitory computer readable medium storing instructions for implementing a method of claim 8 , wherein at least one of the plurality of on chip blocks comprises multiple ports connected to one of (a) multiple routers of the plurality of routers and (b) multiple ports of the one router of the plurality of routers.

12. The non-transitory computer readable medium storing instructions for implementing a method of claim 8 , wherein at least one of the plurality of routers comprises multiple ports, each of the multiple ports connected to a different one of the plurality of on chip blocks.